Holberg

Web service for precise neurological EEG diagnosing

DOMAIN

Healthcare

about the SERVICE

Healthcare web service with rich UI for extensive EEG diagnosing. All the diagnoses are made according to the SCOR standard, which is a unified way of describing the state of the patient in neurology. Data filled by doctors and clinicians are converted into a dynamic PDF report of the patient’s study.

TECHNOLOGY

React.js
TypeScript
Electron
Node.js
Docker
Rest API

Unit testing
Docker
Distributed private node packages
SignalR
Microservices

about the client

Holberg EEG AS is a Norwegian based eHealth company with products for standardised structural interpretation, reporting, and sharing of advanced diagnostic data, and is currently focused on the field of Electroencephalography (EEG).

Client’s brief

Provide doctors and clinicians with the complete tool for EEG diagnosing.
Create the UI that precisely follows the SCORE EEG standard.
Build a fast and reliable UI
Ensure the best possible UX
Create a workflow session for the patient's study with the PDF report as an outcome.
Create a solution for PDF report generation.
Apply real-time updates and synchronization with the third-party EEG reader software.

what are the goals

Provide sophisticated Web UI for neurologists.

Ensure the best possible data integrity and User Experience.

Connect the EEG reader software with the Web UI via real-time communication, so the doctors and clinicians get the fastest way to diagnose the EEG findings.

Make the connection between EEG reader and Web UI seamless via automatic start and closure of the session.

Provide the users with the ability to save all of the study data in a final PDF report.

Implement the ability for various hospitals or organizations to have a custom version of the report based on the available data.

Give the early users ability to provide their feedback on the Web UI.

To improve the existing user experience of collecting patient data and generating reports

Design and build clear and easy-to-use reports and reporting workflow

challenges

Complex neurological data

Complex list pages with pagination, exhaustive filtering, and sorting options.

Autosaving for the complex forms with instant validation according to the SCOR standard.

Some of the forms that heavily rely on the SCORE rules must be implemented with the optimistic updates approach.

API generated SVG Headmodel injection into the UI.

SVG Headmodel interaction.

PDF report generation with dynamic layout.

Real-time communication between the UI and third-party EEG reader software.

Solutions
and Results

SCORE standard rules

Ensure correct implementation. Divide responsibilities across multiple application layers: Databases, API, UI. Data validation on every application layer,data could not proceed if invalid.

Comprehensive UI

Smart sorting and filtering approaches made it possible to create complex filters based on the ability to combine the conditions with AND and OR conditions.

Autosaving process —transparent for the user.

We ensure that auto-saving data is either successfully saved in the DB or rolled back if the request has failed. The request retry strategy was implemented to overcome unstable Internet connections.

Instant action-feedback

If the update on the background fails — data is recovered to the latest successful state and the user gets notified.

Interactive SVG Headmodel

We guarantee safe injection into the UI, after being generated on the API side. To sanitize the contents we had to: List the specific set of allowed SVG tags and attributes to prevent the potential threat or attack. Attach the event handlers, based on the SCORE rules. Apply optimistic updates handlers to all of the possible Headmodel actions.

Flexible PDF report service

Ability to generate different versions of the report layout on demand. Every hospital can get a different layout according to its needs. Or, to even have numerous versions of the layout for maximum flexibility. The content of the report was built with the React components, which were also reused in the main application for the instant Report preview. The report layout is described via JSON file according to the predefined schema blocks.

Real-time updates

Only UI-defined messages are accepted, and unknown types of messages are ignored for security purposes. We used SignalR. The doctors and clinicians now can interact with the third-party EEG reader software and get the updates on the SCORE UI in real-time, no page refresh is needed.